Fraunhofer Institute for Digital Media Technology IDMT
Audio segment matching is a technology that detects identical audio segments/matches among audio items or streams. It can be used for duplicate detection and removal, metadata and rights tracking within and across archives, and to detect content reuse e.g. for media analysis.
Audio phylogeny, on the other hand, is about analyzing the processing history of content, which can be useful for duplicate detection because it helps to identify original versus derived versions e.g. created by transcoding or reencoding within a set of duplicates (in order to decide which version should be kept and which can be safely removed.
Audio Segment Matching
Application Scenarios
Duplicate detection and removal: Audio segment matching can be used to detect partial duplicates creates by partial reuse, cutting or pasting of audio segements during the production process. In order to further support the decision on which duplicate to keep and which duplicate to remove, audio phylogeny can be useful.
Metadata propagation: Audio segment matching is also useful to automatically propagate metadata for reused material during production: Using it, a system is capable of automatically detecting reuse, and to link all reused segement to the same metadata (including rights metadata). This avoids tedious and error-prove manual annotation.
Finally, audio segment matching can be used to detect reuse of material across stations and over time, which can be used for media analysis, e.g. to find out how news reports to a certain topic were presented over time and/or across different broadcasters.
For many A/V content usage scenarios, it makes sense to combine audio and video segement matching.
Licensing Information
Audio Segment Matching is available for licensing, and audio phylogeny will be available for licensing soon – please get in touch with us if you are interested in more information.